Nvidia drivers broken after full upgrade

Hello everyone
So I’m a newbie when it comes to drivers and package management, I’ve had installed nvidia drivers before and it used to work flawlessly. Today I ran full system upgrade (sudo pacman -Syu) after which I’ve noticed laggy graphics - for example youtube videos became very choppy. Also, I’ve noticed that picom has failed to run (I use dwm + picom). When I try to run it manually, it errors:

~ > picom
pci id for fd 8: 10de:1c03, driver (null)
pci id for fd 9: 10de:1c03, driver (null)
pci id for fd 10: 10de:1c03, driver (null)
glx: failed to create dri3 screen
failed to load driver: nvidia-drm
[ 11/04/2025 02:12:24.534 glx_init ERROR ] Failed to enable vsync.
pci id for fd 14: 10de:1c03, driver (null)
pci id for fd 15: 10de:1c03, driver (null)
pci id for fd 16: 10de:1c03, driver (null)
glx: failed to create dri3 screen
failed to load driver: nvidia-drm
zsh: segmentation fault (core dumped)  picom

Here is some info:

~ > uname -a
Linux artyshan 6.17.7-arch1-1 #1 SMP PREEMPT_DYNAMIC Sun, 02 Nov 2025 17:27:22 +0000 x86_64 GNU/Linux
~ > lspci -k | grep -EA3 ‘VGA|3D|Display’
00:02.0 Display controller: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630]
DeviceName: Onboard - Video
Subsystem: Micro-Star International Co., Ltd. [MSI] Device 7b29
Kernel driver in use: i915

01:00.0 VGA compatible controller: NVIDIA Corporation GP106 [GeForce GTX 1060 6GB] (rev a1)
Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3283
Kernel driver in use: nvidia
Kernel modules: nouveau, nvidia_drm, nvidia
~ > pacman -Qs nvidia
local/egl-gbm 1.1.2.1-1
    The GBM EGL external platform library
local/egl-wayland 4:1.1.20-1
    EGLStream-based Wayland external platform
local/egl-x11 1.0.3-1
    NVIDIA XLib and XCB EGL Platform Library
local/libvdpau 1.5-3
    Nvidia VDPAU library
local/libxnvctrl 580.95.05-1
    NVIDIA NV-CONTROL X extension
local/linux-firmware-nvidia 20251021-1
    Firmware files for Linux - Firmware for NVIDIA GPUs and SoCs
local/nvidia-dkms 580.95.05-1
    NVIDIA kernel modules - module sources
local/nvidia-hook 1.5.2-1
    pacman hook for nvidia
local/nvidia-inst 25.10.2-1
    Script to install/uninstall nvidia driver packages in EndeavourOS
local/nvidia-settings 580.95.05-1
    Tool for configuring the NVIDIA graphics driver
local/nvidia-utils 580.95.05-1
    NVIDIA drivers utilities

Did you reinstall the driver by using nvidia-inst tool ?

I first tried reinstalling them manually with sudo pacman -S nvidia nvidia-utils lib32-nvidia-utils but it didnt work, then I installed nvidia-inst and tried installing with nvidia-inst and nvidia-inst –-open, both of which didn’t help also.

Also, I don’t know if it might be related, but I needed to manually reinstall linux-firmware due to https://archlinux.org/news/linux-firmware-2025061312fe085f-5-upgrade-requires-manual-intervention/ (I haven’t upgraded the system for a long time)

The 1060 cannot use open. You need closed.

Also download envycontrol for switching between igpu and dgpu

1 Like

For the record, please show the full output of command

nvidia-inst --test

It changes nothing in your system because of option --test.

Do you have any nvidia-related kernel parameters?

~ > nvidia-inst --test
2025-11-04 17:29:23: Note: 01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GP106 [GeForce GTX 1060 6GB] [10de:1c03] (rev a1) (prog-if 00 [VGA controller])
2025-11-04 17:29:23: Note: Currently installed packages related to Nvidia:
2025-11-04 17:29:24:       egl-gbm 1.1.2.1-1
2025-11-04 17:29:24:       egl-wayland 4:1.1.20-1
2025-11-04 17:29:24:       egl-x11 1.0.3-1
2025-11-04 17:29:24:       libvdpau 1.5-3
2025-11-04 17:29:24:       libxnvctrl 580.95.05-1
2025-11-04 17:29:24:       linux-firmware-nvidia 20251021-1
2025-11-04 17:29:24:       nvidia-dkms 580.95.05-1
2025-11-04 17:29:24:       nvidia-hook 1.5.2-1
2025-11-04 17:29:24:       nvidia-inst 25.10.2-1
2025-11-04 17:29:24:       nvidia-settings 580.95.05-1
2025-11-04 17:29:24:       nvidia-utils 580.95.05-1
2025-11-04 17:29:24: Info: nvidia-inst version 25.10.2-1
2025-11-04 17:29:24: Info: Command line: nvidia-inst --test
2025-11-04 17:29:24: Info: Selected mode: nvidia (Nvidia's proprietary closed source)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
COMMANDS TO RUN:
    ==> Looks like all requested packages and settings are already OK!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Do you have any nvidia-related kernel parameters?

I haven’t messed with it myself, so I assume no. sudo sysctl -a | grep nvidia returns nothing

Out of curiosity… have you tried rolling back to a driver that worked ?

Are you sure it’s the nvidia drivers that got borked ? Have you checked what packages you updated ? Maybe there is something else causing this, because from the output you posted it looks okay.

I would try running an X11 session with these drivers too, just to check if you have the same issues there.

Ok so I’ve checked the previous versions of nvidia, nvidia-utils, lib32-nvidia-utils from pacman log and reinstalled them from the pacman cache. The system then booted into very low resolution and I could not even launch any application, so I assume it’s a fail. Reverted it back to the newest versions, the problem described in my original post remains.

Now I looked into journalctl from the current boot (with the newest drivers) and it looks like nvidia-drm has been loaded properly

Nov 04 22:32:16 artyshan kernel: [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1

but picom still throws

failed to load driver: nvidia-drm

I would try running an X11 session with these drivers too, just to check if you have the same issues there.

I am currently using X11

now that I think about it… I’ve noticed that anytime something more “complicated” happens on the screen, like playing a video or launching a new window, my CPU fan revs a bit, which I haven’t noticed earlier. Could it be that the discrete card is just unused now for some reason, and the CPU usage corresponds to the integrated GPU handling everything?

Update: I’ve installed envycontrol and set the mode to nvidia, but nothing changed

Update2: I’ve checked the usage with nvidia-smi and according to it, the only processes using the Nvidia card are Xorg and wezterm. Memory usage sits at almost nothing, despite launching a webgl stress test in the browser.

Update3: I launched a session on Plasma on wayland, and it seems to be working fine (I originally installed EndeavourOS with Plasma). So it might be an issue with my crappy setup, not with drivers.

Ok, found the solution - apparently new Xorg packages are faulty:

https://www.reddit.com/r/archlinux/comments/1ojfb8r/xorgserver_21120_needs_to_be_reverted_trouble/

after downgrading to 21.1.18-2 all problems are gone.

Good find!

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.